فصلنامه هوش محاسباتی در مهندسی برق
سال دوازدهم شماره 2 (تابستان 1400)

This study optimizes the consumption of electrical energy by monitoring the power consumption caused by the activities of residents at different time intervals during the day and night and stores their electricity consumption in a database to create predicted models based on machine learning methods. Modeling the energy consumption of smart buildings, and then by presenting an algorithm for machine learning based on energy efficiency management system for automatic operation of home appliances based on the previous behavior of residents, leads to the formation of automatic smart building without resident intervention. Managing and monitoring of energy supply and demand process and integration of home solar panels in the building to supply part of the energy consumption was the main advantage of implementing smart grid technology in the building under study. This study showed that 9 kWh of electricity is generated daily from home solar panels. Finally, by comparing each part of the building with a similar normal building in the presence scenario where residents have the highest energy consumption, optimization results were displayed. So that in the lighting system to 25%, the outlet system to 15%, and the cooling and heating system about 40% of energy consumption was saved without reducing the comfort level of residents.

Providing constancy of the nominal voltage level is a major concern in the planning of electric power systems. One of the effective methods to achieve a nominal voltage level is controlling the exciter voltage of the generator. This is done by using AVR. To ensure a better performance of the AVR system under whole operating conditions, the employing of a robust control strategy is essential. In this paper, the PID controller is adopted as a control strategy. Despite its simple structure, the setting of the PID controller parameters is difficult to obtain the robust closed-loop system. In general, there is no straightforward relation between the performance indexes and the controller parameters. To overcome this problem, the control design problem is transformed into an optimization one, and a new performance criterion function is introduced. This new function includes both time-domain and frequency-domain specifications. The Gain and phase margined are maximized in this function to ensure the robustness of the system despite uncertainties; meanwhile, the rise time and overshoot of the step response are minimized to achieve the best performance in the time domain. Ant colony optimization for the continuous domain is used for solving this multi-objective optimization problem. Detailed comparative simulations are carried out via the Monte-Carlo method to illustrate the effectiveness of the proposed control strategy.

Environmental concerns and recent developments in electric vehicle (EV) technology have attracted the attention of the international community to use EVs. In this paper, by considering an urban transportation network, a design strategy is presented to maximize drivers' travel comfort in urban trips. Further, the presented strategy meets various constraints, such as properly locating charging stations in the city, considering traffic volume, taking the shortest route per trip, reducing charging waiting time, etc. The travel comfort index in this paper corresponds to a situation in which the driver does not experience a depleted battery during a trip and successfully finish the trip. Therefore, charging stations should be located throughout the city so that the drivers can access them. Since the movement of vehicles over the course of a day does not follow any particular pattern, in this paper, we use unscented transformation (UT) to investigate the uncertainty in different probabilistic parameters of EVs. Moreover, by clustering the locations of EVs within the urban area, the optimal locations of EV charging stations are determined over the course of the day using an objective function based on a genetic algorithm. The simulation results of the urban transportation confirm the efficacy of the proposed method.

Towards realizing an intelligent networked society, enabling low-cost low-energy connectivity for things, also known as the Internet of Things (IoT), is of crucial importance. While the existing wireless access networks require centralized signaling for managing network resources, this approach is of less interest for future generations of wireless networks due to the energy consumption in such signaling and the expected increase in the number of IoT devices. Then, in this work, we investigate leveraging machine learning for distributed control of IoT communications. Towards this end, first, we investigate low-complex learning schemes that are applicable to resource-constrained IoT communications. Then, we propose a lightweight learning scheme which enables the IoT devices to adapt their communication parameters to the environment. Further, we investigate analytical expressions presenting the performance of a centralized control scheme for adapting communication parameters of IoT devices and compare the results with the results from the proposed distributed learning approach. The simulation results confirm that the reliability and energy efficiency of IoT communications could be significantly improved by leveraging the proposed learning approach.

In this paper, a robust controller is designed for a multi-agent system comprising a leader agent and  followers. Follower agents are supposed to have linear dynamics with time-delay considering uncertainties in state- space coefficient matrices. The dynamics of the leader agent are similar to the followers; however, without any uncertainty. The purpose of the consensus controller is to make the follower agent state variables to track the leader in presence of time-delay and uncertainties. The necessary condition for consensus is the existence of at least a spanning tree in the graph corresponding to a multi-agent system. If the condition is met, a feedback linearization controller is firstly designed for the leader to reach the desired position; then a  robust controller is proposed for each follower to eliminate the effect of model uncertainties by defining the consensus error dynamic between the followers and leader and assuming a specific amount of delay for all agents. High precision, fast convergence, and good robustness against uncertainties are ensuredthanks to the newly proposed control scheme. Finally, asymptotic convergence of the consensus error to zero is guaranteed and LMI conditions for the stability of the closed-loop system is presented. Simulation results show the stability and effectiveness of the proposed controller.

During a power swing, the measured impedance may enter the relay protection zone and cause the relay to operate incorrectly. If the relay operates during power swing, it may cause the network to become unstable. Fault detection during a power swing in compensated lines via a series capacitor is more complicated than simple lines. In compensated lines, the amount of the fault current, the series capacitor and its protection equipment, the type of fault, the fault location, the type of capacitor protection, and the compensation level of the series capacitor are the factors affecting the impedance of the fault loop. In this paper, a new method for detecting fault during power swing is presented by using the magnitude of negative sequence phasor of current with the use of the Taylor series. The proposed method is simulated on the IEEE standard network and its performance has been evaluated. This method has high-speed performance for a variety of symmetric and asymmetric faults. The PSCAD and MATLAB programs are used to simulate the proposed method. Comparing the results with other methods shows the efficiency of the method presented in this paper.

One of the most important parts of infrared search and tracking systems is image processing, which detects the targets in infrared images. In this paper, a new infrared small target detection algorithm is proposed. The proposed method uses heuristic optimization algorithms to find small targets in infrared images. In this way, the particle swarm optimization algorithm is used as one of the best heuristic optimization algorithms. The performance of the proposed algorithm is evaluated using real and simulated infrared images. These images include a variety of false response sources like high-intensity edges, cloudy background, complex sea-sky background, and a target close to high-intensity background clutter. The experimental results are compared with the four common methods of infrared small target detection. The experimental results show the effectiveness and performance of the proposed algorithm. Also, the runtime of the proposed method is comparable to other commonly used methods, and this makes it usable in real-time programs.

O Optimization is a process that has long been considered by researchers in various fields and its methods have been utilized to increase productivity while reducing costs. Moreover, it is essential to achieve simultaneously several contradictory goals. In control engineering, optimizing controller parameters to attain several different goals could be considered as a significant challenge. In this paper, a fractional-order PID controller is designed for a training helicopter with two DC motors, called Twin-Rotor Multi Input Multi Output System (TRMS). The design of this controller is based on the optimization of different objective functions with metaheuristic multi-objective optimization algorithms. Finally, their Pareto front and different evaluation criteria such as Spacing Metric and Non- uniformity of Pareto Front are employed to compare the performance of these algorithms for optimization of controller parameters on the decoupled TRMS system. The results show the superiority of the proposed multi-objective fractional-order PID controller for the TRMS system.